OpenEdition character set experiment
Rick Troth <troth@casita.houston.tx.us>
BMC
 
 
 
## ) Copyright 1996, Richard M. Troth, (see below).	     <plaintext>
 
 
This is the README file for the OECS demo/experiment.
The point is that a function can be written that automatically handles
either an ASCII or an EBCDIC input stream  (assuming the input is textual)
and returns NULL-terminated strings of the  correct local character set.
Such a function can be used in shells, compilers, and any text processing
components in a POSIX system like IBM's OpenEdition VM or OpenEdition MVS
to facilitate interoperability with systems using the "foreign" base
character set.   Use of such a function for reading input streams is
expected to reduce effort on the part of the customer,  which will likely
result in better sales of the system.
 
Remember:  interoperability is more important than compatibility.
We lose the brass ring that OpenEdition offers if we don't surmount
problems like this.   ASCII toleration is only the first step.
There's more work to do.   VM,  especially now with OpenEdition,
provides an operating environment that is truly unique in the
software industry.   Let's keep going!
 
 
After you've extracted  oecs.tar  (and converted all sources
except oecstest.sh into EBCDIC if on OEVM),  do a
 
		make
 
and then
 
		oecat some_ASCII_file
 
and/or
 
		oecat some_EBCDIC_file
 
The output of either should be readable on your screen.
Foreign shell scripts can often be executed with
 
		oecat some_script | sh
 
(shell scripts that read from standard input excluded here)
Like maybe
 
		oecat oecstest.sh | sh
 
The  'oecat'  program will automatically sense ASCII or EBCDIC input,
convert lines as they are read,  and write the native character set
to its output stream,  rendering foreign files readable and native files
unharmed.   Oecat is something that you might place in  /usr/local/bin.
 
 
The  getline()  function is where the real logic for this operation
is found.   Yes,  I know that this  getline()  is horribly inefficient
for a wide range of systems.   But it compiles and runs as intended
on a wide range of systems.
 
 
Permission is granted to use the files found in this directory
for  experimentation,  education,  and inclusion in software for sale.
The author,  Richard M. Troth,  and any employers,  employees,
associates,  or relatives thereof,  shall  NOT BE HELD LIABLE
for any damages resulting from the use or sale of this software.
This software package is provided as-is in good faith
for the common good and advancement of the art.
 
 
The files in this package are
 
	README (this file),
	makefile, oecat.c, aecs.c, aecs.h, getline.c, putline.c,
	oecstest.sh
 
Note that this last file,  oecstest.sh,  contains a mixture of
ASCII and EBCDIC  (thus  "test"  in its name).   It should be left
untranslated.   The rest of the files are textual sources which
should be translated into EBCDIC when ported to OpenVM.   (in the
general case,  we don't expect to find source files with *both*
ASCII and EBCDIC,  but it makes a good test case)
 
Updates can probably be found at
 
	    http://casita.houston.tx.us/~troth/software/oecs.tar
 
	or at
 
	    http://ua1vm.ua.edu/~troth/software/oecs.tar
 
 
Many thanks to Dave Rivers at SAS for help in getting this to work.
Thanks too to all the Workshop VMers who  "hammered on it".
 
 
--
Rick Troth <troth@casita.houston.tx.us>, Houston, Texas, USA
http://casita.houston.tx.us/~troth/